Hi Yudhayaja, This is more zoomed view. Regards, Yasara On Fri, Jan 22, 2016 at 4:39 PM, Yudhanjaya Wijeratne <[email protected]> wrote:
> Thanks for the update, Yasara. Due to the very small scale that I used the > human line may appear almost horizontal to 0. I noticed every other > algorithm used larger figures. So we may not be seeing the variations > (relatively). Could you use a 5x multiplier and let me know what happens? > > Thanks, > Yudha > > On Fri, Jan 22, 2016 at 4:34 PM, Yasara Dissanayake <[email protected]> > wrote: > >> Hi, >> Graph compare with human manual read rates. >> >> Regards >> Yasara. >> >> On Fri, Jan 22, 2016 at 4:26 PM, Yasara Dissanayake <[email protected]> >> wrote: >> >>> Hi Yudhanjaya, >>> This the sentiment results[1][2] for above articles using following >>> sentimental algorithms >>> >>> 1. Using Naive Bayes Algorithm >>> 2. Standard sentimental analysis: Using positive negative word bucket >>> 3. Afinn sentimental analysis: Using positive negative value for >>> each word >>> 4. Stanford NLP in 3 forms >>> >>> *Stanford Sentiment Method A:Predict each sentence positivity , >>> negativity and getting summation.** And for score use deducting 2 from >>> originally return sentiment score from library* >>> >>> Score for each sentence=-2: "Very Negative" >>> >>> Score for each sentence=-1: "Negative" >>> >>> Score for each sentence=0: "Neutral" >>> >>> Score for each sentence=1: "Positive" >>> >>> Score for each sentence=2: "Very Positive" >>> >>> *Stanford Sentiment Method B:Use main sentiment as longest sentence.And >>> took Rank as longest_sentence_score*length_of_longest_sentence* >>> >>> *Stanford Sentiment Method C: **Predict each sentence positivity , >>> negativity and getting summation. And for score use originally return >>> sentiment score from library* >>> Score for each sentence=0: "Very Negative" >>> Score for each sentence=1:"Negative" >>> Score for each sentence=2:"Neutral" >>> Score for each sentence=3:"Positive" >>> Score for each sentence=4: "Very Positive" >>> >>> Regards, >>> >>> Yasara. >>> >>> >>> >>> >> > > > -- > Yudhanjaya Wijeratne > Marketing Officer, WSO2 Inc <http://www.wso2.com> > +94775496911 | @yudhanjaya >
_______________________________________________ Dev mailing list [email protected] http://wso2.org/cgi-bin/mailman/listinfo/dev
